Geophysical and geologic study of the sub-surface structure of the earth continues to be a important field of endeavor for several reasons. The continued search for producing reservoirs of hydrocarbons, such as oil and gas, as well as for mineral deposits and sources of geothermal energy, is a particularly important motivation for obtaining information about the earth's crust. This information is also important in monitoring the movement of plates of the earth crusts, and thus in predicting and measuring geologic events ranging from slips which create hydrocarbon reservoir traps to catastrophic earthquakes. As a result of the economic importance of this technology, significant effort continues to be expended in developing new and improved surveying and interpretation techniques.
Conventional seismic surveying is generally performed by imparting energy to the earth at one or more source locations, for example by way of a controlled explosion, mechanical impact, or the like. Return energy is then measured at surface receiver locations at varying distances and azimuths from the source location. The travel time of energy from source to receiver, via reflection from and refraction by sub-surface strata and interfaces therebetween, is indicative of the depth of the strata and interfaces. In addition, various attributes of the detected energy (e.g., the difference in velocity of pressure versus shear waves, phase differences between input and received energy) are indicative of the composition of the strata.
This time-domain seismic information is conventionally interpreted to generate a survey of the sub-surface geology for the region of interest, usually in the form of a contour map indicating the location, depth and acoustic velocity of various sub-surface strata. From such maps and other presentations of the data, skilled geologists and geophysicists can infer the location and depth of potential hydrocarbon reservoirs, and other structures such as faults, strike-slips and the like at which reservoir traps may be present.
While conventional analysis of such seismic survey data is successful to a large degree, certain inaccuracies are often present. The detected signal will contain a "signal" portion corresponding to the reflected or refracted source energy. A "noise" component will also generally be present in the detected signal that masks the signal and reduces the sensitivity of the survey to strata and interfaces; such noise can be generated by the signal itself, or may be generated by external effects and thus unrelated to the signal. In addition, poorly designed surveys can result in spatial aliasing and related undesirable effects which cloud the interpretation. Furthermore, inaccuracies in the estimated acoustic velocity of particular formations, or in other assumptions used in data interpretation, will result in error in the resulting survey or contour map. According to conventional techniques, it is not uncommon for errors in the position or depth (or both) of a particular geologic structure to have magnitudes in the hundreds of feet.
Particularly in the case of drilling hydrocarbon wells, the cost of physically verifying the seismic survey, or alternatively the cost of a "dry hole" if the survey is inaccurate, can be quite significant. Furthermore, the margin for error in the seismic survey is continuing to shrink, as the difficulty in locating a producing well has increased in recent years due to prior exploitation of the more easily located reservoirs, with the remaining deeper and more difficult reservoirs being those to which current survey techniques are directed. In addition, particularly as the depth requirement for seismic surveys increases, the signal-to-noise ratio of the seismic energy becomes poorer. As a result of all of these factors, the modern exploration geophysicist is required to perform a task of increasing difficulty (accurate seismic analysis of deep and small reservoirs) with poorer quality information.
Furthermore, the cost of obtaining field data for seismic surveys has remained high over recent years. The analysis of previously existing seismic data (i.e., source and receiver traces) using new interpretation techniques has therefore become more popular in the industry, eliminating the cost of obtaining new seismic field data. Considering that several competitors may be analyzing the same raw data, proprietary data analysis routines and techniques have become an important asset of the exploration company. Significant development activity of new and improved noise reduction and cross-correlation techniques has resulted, in attempts to improve the resolution and sensitivity of the surveys, so that deeper and smaller sub-surface strata and interfaces can be located. Since the actual structure of the earth beneath any given location is not fully known, evaluation of new data analysis and processing techniques is quite difficult, and is limited to the comparison of new results against prior surveys or, if a newly discovered formation is indicated, by core sampling, drilling or the like.
Heretofore, new seismic data analysis routines have been tested and verified by way of models of the structure of a portion of the earth. These geophysical models, both of the numerical and physical types, have generally been designed to match actual regions of the earth being explored. Construction of a model of the surveyed region according to a contour map, followed by the performing of a scaled "seismic" survey on the model, provides a comparison of the seismic data acquired from the model with that from which the contour map was constructed. Differences between the field seismic data and that from the model indicate inaccuracies in the seismic data interpretation process originally applied to the field data. Re-interpretation of the previously acquired seismic data, or even acquisition of new seismic data from the region of interest, can then be performed to provide a new or adjusted survey of sufficient accuracy that drilling could be performed with reasonable confidence.
For such geophysical modeling, scale models of the earth have been formed by way of plaster and wood molds designed to match shapes in the contour map. The material molded by such molds is generally one of several two-part rubber or plastic materials having the desired physical properties, such as acoustic velocity (compressional, horizontal and vertical shear), density, and other elastic material properties. Each molded layer becomes a portion of the mold for the next adjacent layer, with the result being a laminated block of dimensions on the order of one to three feet on a side to represent the surveyed region. Scaled acoustic or ultrasonic sources and detectors are then deployed at the surface of the model, generally near the center of the top surface so that boundary effects at the sides of the model are effectively infinitely distant, and a scaled seismic survey is performed to simulate an actual field survey.
However, these prior scale models have not provided sufficiently accurate information on a timely enough basis to allow for useful verification of the survey information. This is due to the time-consuming and expensive construction of the molds, such that the use of the model is seldom sufficiently timely to meet business needs. Furthermore, the precision of this fabrication technique is limited to on the order of tenths of inches; for a typical scale of 1 inch to 1000 feet, an error on the order of tenths of inches results in a deviation of on the order of hundreds of feet in the earth. Therefore, not only are such simulations late, the results are also insufficient to determine if a deviation in the data is due to inaccuracy in the scale model, or truly due to inaccurate interpretation of the field data.
By way of further background, copending U.S. application Ser. No. 714,272, filed Jun. 12, 1991, incorporated herein by reference, and assigned to Atlantic Richfield Company, describes a new method for fabricating a scale model using stereolithography. This method provides a smaller and more accurate geophysical model, with much reduced fabrication time, as compared with the conventional plaster molding technique described hereinabove. The precision of a model formed according to stereolithography, as described in this copending application, is much superior than such conventional methods.
Actual measurement of the finished scale model would allow one to account for dimensional inaccuracy of the model in analyzing the simulation data. However, actual measurement of the model must be performed in a non-contact manner so that the model can be useful after such measurement. Conventional non-contact measurement techniques, such as x-ray, CAT scan, or other imaging techniques, are not only expensive, but are quite cumbersome for objects of the size of these models, generally on the order of two to three feet on a side for the plaster molded models; such imaging techniques, while more convenient for the scale models on the order of one foot or less for the models formed according to application Ser. No. 714,272, are still inconvenient, at best.
An alternative method for measuring a completed scale model is to use a scaled seismic survey, applying acoustic energy to the model at a surface location, and detecting reflected energy at surface receiver locations of the model. In order to generate an accurate zero-offset trace from such techniques, however, significant data processing is required, particularly considering sixty or greater "fold" of data per midpoint, as is conventionally used for such surveys. In addition, the use of a seismic survey method to measure the model is not appropriate where the seismic survey method itself is to be ultimately measured using the model. Furthermore, the angles at which the acoustic energy may be reflected, especially considering the finite boundaries of the scale model, can cause interference in the detected vibrations, which carl cloud the results such that accurate measurement of the scale model may not result.
It is therefore an object of the present invention to provide a method and system for measuring the actual structure of a scale model of the earth's crust.
It is a further object of the present invention to provide such a method and system which may be performed efficiently, and in a non-destructive or non-invasive manner.
It is a further object of the present invention to provide such a method and system for obtaining zero-offset traces of the scale model of the earth, for use in development and verification of new and existing seismic data analysis techniques.
Geologic modeling, as opposed to geophysical modeling described hereinabove, simulates time-dependent movement of layers in the earth crust by stressing portions of a scale model and monitoring the response of the model to the stress stimulus. Such modeling is useful in understanding the effects of stress and strain on sub-surface structures, for purposes of hydrocarbon exploration (as reservoir traps are often created by such forces) and for predicting and measuring natural seismic activity. Conventional scale geologic models are formed of non-elastic materials which are somewhat deformable by physical pressure; examples of such materials include clays and dry sand. The scale model of the earth is formed in layerwise fashion using these materials, for example in a "sandbox" or other container. Generally, initial conditions such as faults, inclusions and the like are built into the model.
The simulation of the response of the earth to stress and strain forces is a time sequence event, generally lasting on the order of one hour or less, in which a known force (both direction and magnitude) is applied to a side of the model or a particular layer or layers therein. Monitoring of the model before, during and after the application of the force provides information regarding the behavior of the modeled structure to the model force.
Conventional techniques include visual or photographic inspection of the model surface before and after application of the force, from which the geologist may be able to infer the sub-surface response to the force. Sub-surface information is conventionally obtained from the model by immobilizing the model and slicing vertically or horizontally therethrough to reveal cross-sections of its internal structure, for example cross-sections at vertical "dips". Sand models generally require some amount of solidification prior to sectioning. Examples of such techniques are described in Naylor, et al., "Fault geometries in basement-induced wrench faulting under different initial stress states", J. Structural Geology, Vol. 8, No. 7 (Pergamon, 1986), pp. 737-752 (sectioning of a sand model after impregnation with gelatin), and in Hildebrand-Mittlefehldt, "Strain fields in and around boudins in a clay experiment", J. Structural Geology, Vol 5, No. 3/4 (Pergamon, 1983), pp. 465-470 (sampling of a clay model after replacement of water with a plastic). CAT scans have also been performed of the stressed scale model, if the size of the model is limited (e.g., for models having a thickness of on the order of up to one foot).
These conventional measurement techniques for geologic modeling present significant limitations to the analysis of the model results, however. While surface inspection provides real-time data in a relatively easy manner, such inspection provides little direct information concerning sub-surface response to the applied force. While sectioning or core sampling of the model provides the desired information about the internal model structure after stress, this information can only be acquired on a sampled basis, such that important sub-surface response may not be detected at locations missed by the samples and sections. Furthermore, the process of sectioning or sampling, and analysis of the sections or samples to infer the response of the entire model, is quite time-consuming.
Furthermore, significant information regarding the geologic modeling, particularly the dynamics of structures internal to the model, cannot feasibly be obtained using conventional sectioning and sampling techniques. This is because the sectioning of the model is not only destructive at the sectioned location, but requires relief of all stresses prior to sectioning (to avoid the sectioning itself from relieving the stress). Furthermore, as noted hereinabove, additional processing to solidify or otherwise stabilize the model structure is often necessary. Accordingly, sectioning of the model necessarily is performed at the end of the stress event, as continued stress after the sectioning would not provide an accurate model of the earth response.
In particular, geologic phenomena such as the type and amount of displacement and deformation are highly time-dependent, with the time-dependency often being quite complex. For example, a small fault may initially slip as the force is applied, remain inactive for some time, and become active again later, splaying into several faults. Such complex dynamics may not be visible at the surface and therefore not detectable by real-time visual inspection, and of course are not directly detectable by sectioning or sampling at the end of the test in the conventional manner. Other time-dependent dynamics useful in determining the cause of folding, compaction (i.e., dewatering) and other geologic effects also cannot be monitored merely by end-of-test sectioning or sampling.
Time dependent monitoring of the model by cross-sectioning could conceivably be performed in an iterative fashion, with sections made at varying amounts of time or force. However, such iterative modeling would require the rebuilding of the model for each iteration, as the test cannot be continued after sectioning due to destruction of the model and relief of existing stresses and strains. Such reconstruction of the model would not only be time-consuming and expensive, but the size, elasticity and other attributes of the modeled structure will necessarily vary from iteration to iteration, incorporating error into the iterative results.
It is therefore an object of the present invention to provide a method and system of geologic modeling where real-time information concerning sub-surface structures can be acquired during the application of force to the scale model.
It is a further object of the present invention to provide such a method and system where the information is acquired in a non-destructive manner.
It is a further object of the present invention to provide such a method and system useful both with clay and sand geologic models.
Other objects and advantages of the present invention will be apparent to those of ordinary skill in the art having reference to the following specification together with its drawings.